ChatGPT fixes ‘big’ problem; CEO calls change ‘small but…’; addresses user frustrations

OpenAI has addressed ChatGPT’s notorious overuse of em dashes, allowing users to instruct the chatbot to avoid them via custom settings. This fix, arriving with the GPT-5.1 model, ends years of user complaints and highlights the challenges in controlling AI language models, suggesting true artificial general intelligence may be further off than anticipated.